And in deal.ii, I use the weak form:
weak.JPG <about:invalid#zClosurez>
solve and plot F (which is neither right, nor continuous ):
On of the problems is that the exact F is not continuous -- the arctan has a
discontinuity along the negative x-axis. So if you try to project a
discontinuous function onto a continuous space, you will always get oscillations.
What I would try in your case is to first see if you can use the *exact* phi.
I.e., don't project phi into a finite element space first, but instead compute
the exact function F (which here you actually know) and then project that into
a finite element space for the function F_h. For F_h, you can either use
FE_Q(1)^dim or FE_DGQ(1)^dim as an experiment.
I don't know whether that will be accurate enough for your purposes (it will
still have oscillations if you use FE_Q), but it's good practice when
debugging to try and isolate problems to one issue at a time -- so use the
exact F to project onto an F_h, rather than first project phi onto phi_h and
then arg(grad phi_h) onto an F_h. That's just too many steps at once where
something can go wrong.
Cheers
W.
--
------------------------------------------------------------------------
Wolfgang Bangerth email: [email protected]
www: http://www.math.colostate.edu/~bangerth/
--
The deal.II project is located at http://www.dealii.org/
For mailing list/forum options, see
https://groups.google.com/d/forum/dealii?hl=en
---
You received this message because you are subscribed to the Google Groups "deal.II User Group" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
For more options, visit https://groups.google.com/d/optout.
